DESCRIPTION (adapted from the application) Lower extremity ulcers are a serious complication of diabetes mellitus. More than 16 million people in the US have diabetes mellitus and 15% of them can expect to develop a foot ulcer at some point in their life. Annually more than 50% of all non-traumatic amputations occur in patients with diabetes, thereby making diabetes the leading cause of lower extremity non-traumatic amputation. Lower extremity chronic wounds precede more than 85% of these amputations. In diabetics, the etiologies of these wounds are believed to be lower-limb arterial insufficiency, neuropathy, or a combination of both. The patients with diabetic neuropathy are often managed medically while those with arterial insufficiency are treated by vascular surgical intervention. Previous studies have shown that both poor lower-limb arterial blood flow and diabetic neuropathy are associated with the risk of developing a foot ulcer and eventually an amputation. However, very little has been published on the risk factors or prognostic factors associated with the failure of a patient with a diabetic neuropathic foot ulcer to heal. This is problematic, since new medical treatments recently approved by the FDA were specifically labeled for the treatment of diabetic neuropathic foot ulcers. Without knowledge of risk and prognostic factors, it is difficult for health care practitioners to make informed decisions with respect to whom they should treat with standard care and it is very difficult for a clinical investigator to plan well designed clinical trials. Using the largest wound care specific database and multivariable regression techniques, we will conduct a series of cohort studies to create explanatory and prognostic models. The explanatory models will be used to estimate the association of a risk factor on the likelihood that a wound will heal (or require an amputation) by the 20th week of standard therapy. The prognostic models will be used to estimate the probability that an individual with a neuropathic foot ulcer will heal with standard therapy. Ultimately, parsimonious clinically friendly models will be developed from complex models, so that a healthcare provider can discriminate between those wounds that will heal (or require an amputation) with standard care and those wounds that will not heal (or require an amputation). These models will not only be useful to health care providers, but they will also be useful to clinical investigators trying to design clinical trials on patients that might maximally benefit from a new treatment.